Comments on anatomy and temporal ontologies

Sun May 19 02:02:57 EDT 2002

I am broadly supportive of the TAIR comments. I would only express caution
about the opinion on whether or not terms should be introduced prior to
use... I've been approached by barley, wheat and oats people about getting
involved in ontology building. We can probably get them actively
contributing for these grains, if we encourage it.

Cheers
Richard at IRRI

-----Original Message-----
From: Tanya Berardini [mailto:tberardi at acoma.Stanford.EDU]
Sent: Saturday, May 18, 2002 12:18 AM
To: Vincent, Leszek
Cc: Curators At TAIR (E-mail); Doreen Ware (E-mail); Ed Coe (E-mail);
Gramene developers; Lincoln Stein (E-mail); Mary Polacco (E-mail);
Pankaj Jaiswal (E-mail); Richard Bruskiewich (E-mail); Susan McCouch
(E-mail); Victor Ulat (E-mail); Leszek Vincent; Michael Ashburner
(E-mail); SPRIG-CVO_Development (E-mail)
Subject: Comments on anatomy and temporal ontologies

I have summarized specific comments from the TAIR group below. As you will
see, they are quite extensive yet arise from only a partial review of the
ontologies.  We find that there are many problems that need to be resolved
before the ontologies can be made public. It would be helpful if we can
discuss the content issues by email: discussions of definitions, what
should be term synonyms, the goal of a generic temporal framework that can
accomodate all plants, etc.

What should our next step be?  What are your opinions on going back to
the basics vs. trying to fix the current ontology?

Tanya

*********

ANATOMY:

(1) Wrong definitions, duplicate terms, need for synonyms, removal of
TAIR terms.

	Example:

	ptyxis PO:0005108
	Definition: leaves in an embryo
	Synonym: embryonic leaf

	However, upon further investigation, I find the following defs for
	this term:

	Ptyxis: The rolling and folding of embryonic leaves in relation to
	a central point, transverse axis or longitudinal axis; a
	combination of transverse and longitudinal posture for embryonic
	parts

	Ptyxis:The way in which a leaf is sometimes folded in the bud.

	Embryo leaves (ptyxis)

	The folding of mature leaves is called ptyxis. Two common
	types of ptyxis are:
	Circinate: The apex of the leaf is curled towards
	the base. Many ferns exhibit circinate vernation in their
	"fiddleheads."
	Revolute: The margins of the leaf are curled under. An
        example is labrador tea (Rhododendron groenlandicum).

	Oddly, there is a child of ptyxis that is called vernation
	which is defined as classification based on rolling or folding of
	individual embryonic leaves and arrangement of embryonic leaves
	within a structure (synonym ptyxis type)

	So, which is the correct definition and, if it is embryo leaves,
	shouldn't this just be a synonym for cotyledon?

At the same time, there are also the terms:

	Plumule PO:0005347
 	def. Embryonic leaves in seed derived from epicotyl.
	(again, looks like a synonym for cotyledon, also the leaf derives
	from the meristem, not the epicotyl)

	Cotyledon PO:0005358
	def. Embryonic leaf or leaves in seed

	Cotyledon PO:0005519
	def. Embryonic leaf

On the other hand,

	cotyledon PO:0000138 (was TAIR:0000138)
	definition: Leaf or leaf-like organ formed during embryogenesis.
	Specialization includes accumulation of storage reserves that are
	mobilized during seed germination.
	definition_reference: TAIR:lr

is _gone_.

Similarly,

	hypocotyl PO:0000310 (was TAIR:0000310)
	definition: the axial part of the embryo or seedling located
	between the cotyledon or cotyledons and the radicle.

is gone and has been replaced, though not in all the places where
PO:0000310 was
by

	hypocotyl PO:0005342
	def:embyronic stem in seed, located below cotyledons.
	ALTERNATIVELY,
	the axial part of the embryo or seedling located between the
	cotyledon or cotyledons and the radicle.

(2) More duplication examples:

	(a) 2 terms for lamina

"lamina" PO:0000221 (was TAIR:0000221) child of cotyledon
def: Blade,or expanded part of the leaf

"leaf lamina" PO:0005323 child of leaf
def: The blade or flat part of a leaf (GR:pj37)

	(b) 2 terms for "leaf margin" (exact match)

leaf margin PO:0000218 (was TAIR:0000218) child of lamina
no definition

leaf margin PO:0005578  child of leaf lamina
def: the margin of the leaf. also determines shape
of the leaf (GR:pj37)

	(c) 2 terms for hydathode (exact match)

hydathode PO:0000323: (was TAIR:0000323) only partof lamina (i.e. only
cotyledon)
def:  A structural modification of vascular and ground tissues,
usually in a leaf, that permits the release of water through
a pore in the epidermis; may be secretory in function.

hydathode PO:0005660 isa vascular bundle
def: A structural modification of vascular and ground tissues,
usually in a leaf, that permits the release of water through
a pore in the epidermis; may be secretory in function.

- this one raises multiple problems when examing all of its
appearances in the ontology, the true path rule does not
hold

	(d) 2 terms for midvein

"midvein" PO:0000235 (was TAIR:0000235) partof lamina
def: the central or usually most prominent vein in a
leaf or leaf-like organ

"leaf midvein" PO:0005099 isa leaf vein
def: the central conducting and supporting structure of
the blade of a leaflet or a compound leaf

	(e) radicle, same problem

**This seems prevalent.  Can a check for duplications be done
computationally?
You can't tell a priori which terms have duplicates and which don't.

(3) Definitions are sometimes incomplete or meaning less.

	eg., infloresence cyme is defined as : the central flower opens
first.

	according to 'How to Identify Plants' by Harrington

	"Cyme is a flower cluster, often convex or flat topped, in which
the
	central or terminal flower opens first."

However, there is also:

	Cymose infloresence (one that bears cymes) which is a TYPE
	of inflorescence..... this is the pattern of the infloresence of
	something like Queen Anne's lace, wild carrot.

(4) General organization needs to consider true path for all terms! There
are many, many violations of this, for example:

	(a) all of the parts of the generic embryo belong then to the
specific
instances, right now, upper tier and lower tier are parts of the
generic embryo, and are probably not relevant to maize embryos at any
stage of development.

	(b) epicotyl now has coleoptile as a part- fine for grasses but
not a
part of the epicotyl in Arabidopsis

	(c) Not all infloresences have an epicalyx

	(d) floret (sensu Poaceae) a part of an infloresence and not an
instance of a flower

	(e) Then there  is 'radicle type' and that coleorhiza is part of
the  new radicle term... Arabidopsis does not have coleorhizas...  so can
PO:0005357 be made (sensu Poeaceae)

	(f) epiblast (undefined in the ontology PO:0005345) which is
(according to Esau- a small structure opposite the suctellum present
in the embryo of some grasses) is an instance of an embryo- this does
not make sense... this should be a part of relationship to some
instances of an embryo.

	(g) hydathode - see above

(5) Need to get rid of "X type," in favor of having the children be
instances of X: There are _31_ terms that are of the form X_type. For
example:

	(a) embryo type

Instead of

%embryo
.%embryo type
..%embryo bent
..%embryo broad
..%embryo capitate
..etc.

Should be:

%embryo
.%embryo bent
.%embryo broad
.%embryo capitate
.etc.

	(b) fruit type

Instead of

%fruit
.%fruit type
..%fruit accessory
..%fruit aggregate
..etc.

Should be:

%fruit
.%fruit accessory
.%fruit aggregate
.etc.

Related issue:  Do we need to add terms if we are not
going to use them yet?  For example, do we need all of the embryo
types to describe the embryos of rice, maize, barley, sorghum, oat,
wheat and arabidopsis or just a few?  We should think about adding
terms only as we need them and not just because they exist.

(6)  Correct use of sensu

	From the GO site:

	"GO nodes should aggressively avoid using species-specific
	definitions. Many functions, processes, and components, however,
	are not common to all life forms. Our current convention is to
	include any term that can apply to more than one taxonomic class
	of organism.

	Within the ontologies,there are cases where a word or phrase has
	different meanings when applied to different organisms. For
	example, mating in yeast is very different from mating in mice.
	Such terms are distinguished from one another by their definitions
	and by the sensu designation, as in the term mating (sensu
	Saccharomyces). Using the sensu reference makes the node available
	to other species that use the same process/function/component.

	A GO node should never be more species-specific than any of its
	children. Child nodes can be at the same level of species
	specificity as the parent node(s), or more specific. When adding
	more species-specific nodes, curators should make sure that
	non-species-specific parents exist (or add them if necessary)."

I think we should apply similar rules to POC.

In this case, the existing terms:

rachilla (sensu inflorescence)
rachis (sensu inflorescence)
stipe (sensu flower)

do not make sense.

Note that the terms rachilla, rachis, and stipe _also_ exist as
separate PO terms.

Also, Genus and Family names should be capitalized. (i.e. sensu Zea,
sensu Magnoliophyta, etc.)

TEMPORAL FILES:

(1) We should try to not divide these by monocot and dicot but instead,
like Leonore suggested, by actual stages:

sporophytic(new)
 embryonic
 vegetative
 reproductive
 senesence (new)
gametophytic(which would overlap with sporophytic reproductive stages,
e.g. ovule developmental stages)

This will help achieve our goal of using single terms to describe common
processes.

(2) Duplications and definitions (again): There are DBXrefs for terms
like germination PO:0007051 but NO text. So there is germination
PO:0000391 (Arabidopsis- also undefined), PO:0007053 and probably
more.  Can we define germination broadly as including imbibition,
through radicle emergence and then make sensu terms. We can have a
generic germination and then stages according to the organism ... but
they should be developmental stages- NOT processes that are being
defined.

----------------------------------------------------------------------------
---
Tanya Berardini, Ph.D.			tberardi at acoma.stanford.edu
The Arabidopsis Information Resource	FAX: (650) 325-6857
Carnegie Institution of Washington	Tel: (650) 325-1521 ext. 325
Department of Plant Biology		URL: http://arabidopsis.org/
260 Panama St.
Stanford, CA 94305
----------------------------------------------------------------------------
---